Tracking Polymicrobial Metabolism in Cystic Fibrosis Airways: Pseudomonas aeruginosa Metabolism and Physiology Are Influenced by Rothia mucilaginosa-Derived Metabolites

Pseudomonas aeruginosa is a dominant and persistent cystic fibrosis pathogen. Although P. aeruginosa is accompanied by other microbes in the airways of cystic fibrosis patients, few cystic fibrosis studies show how P. aeruginosa is affected by the metabolism of other bacteria. Here, we demonstrate that P. aeruginosa generates primary metabolites using substrates produced by another microbe that is prevalent in the airways of cystic fibrosis patients, Rothia mucilaginosa. These results indicate that P. aeruginosa may get a metabolic boost from its microbial neighbor, which might contribute to its pathogenesis in the airways of cystic fibrosis patients.

C ystic fibrosis (CF) patients experience persistent polymicrobial colonization of their airways. Rothia mucilaginosa and Pseudomonas aeruginosa are microbes frequently detected in CF patient airways, and their cooccurrence has been observed in CF patient sputum (1)(2)(3)(4). Microbes within polymicrobial infections display complex interactions, such as metabolite cross-feeding (5). For example, P. aeruginosa inefficiently metabolizes host-derived mucins. Rather, P. aeruginosa utilizes mucin degradation products from oral anaerobes to support its growth (6,7). Still, many studies of CF-associated microbes are conducted under artificial conditions that fail to take into account the nutrient and oxygen gradients found in CF patient airways (8,9,25). The lack of overlap between laboratory conditions and CF patient airways is reflected by the differences in growth rates, with estimates of bacterial doubling times being 100-fold times lower in sputum than in standard medium (10). Furthermore, most CF studies focus on single microbes. One primary reason for this is the lack of a robust model to examine the microbial interactions. Stable-isotope-assisted metabolomics analyzes the fate of heavy atoms from stable-isotope-labeled precursors to products, which makes it a suitable approach for monitoring metabolites produced by one microbe when cross-fed to a second microbe. In order to further explore cross-feeding interactions between two CF microbes in a relevant environment, we cross-fed labeled glycolysis products from R. mucilaginosa to P. aeruginosa (8). Both strains were isolated from the sputa of CF patients. We believe that our P. aeruginosa strain is representative of CF strains, as its core genome is similar to that of P. aeruginosa strain PA17 and other CF isolates (11). In an effort to mimic the CF airway environment, R. mucilaginosa was fed labeled glucose in anaerobic and aerobic artificial-sputum media, and the R. mucilaginosa supernatant was fed to P. aeruginosa in nutrient-rich (artificial-sputum medium) under low-nutrient (M9 minimal medium) conditions. As P. aeruginosa lacks some glucose utilization capacities, including a key enzyme involved in glycolysis, phosphofructokinase, we postulated that cross-feeding metabolites from R. mucilaginosa impacts the metabolism of P. aeruginosa (12).
R. mucilaginosa metabolism under aerobic and anaerobic conditions. R. mucilaginosa was grown aerobically and anaerobically in artificial-sputum medium (see Text S1 in the supplemental material). Under both anaerobic and aerobic conditions, glucose-derived 13 C was incorporated into glycolysis metabolites, namely, lactate and acetate, and some amino acid biosynthesis pathways in R. mucilaginosa ( Fig. 1; Table S1). The labeled glucose was not incorporated into the tricarboxylic acid (TCA) cycle, pentose phosphate pathway, or long-chain fatty acid biosynthesis pathways. For most FIG 1 Glucose-derived 13 C was incorporated into pyruvate, lactate, acetate, alanine, valine, serine, glycine, leucine, and isoleucine in R. mucilaginosa under both anaerobic and ambient-oxygen conditions. Mϩ2, Mϩ3, Mϩ4, and Mϩ5 indicate compounds that contained 2, 3, 4, and 5 13 C atoms, respectively. Isotope enrichment means an abundance of labeled ion/unlabeled ion (corrected for natural abundance). Isotope enrichment was greater at 24 h than at 8 h or 4 h. For pyruvate, alanine, valine, and acetate, greater isotope enrichment was observed under anaerobic conditions at 24 h. For lactate, glycine, serine, and isoleucine, greater isotope enrichment was observed under ambient-oxygen conditions at 24 h. The incorporation of glucose-derived 13 C into leucine biosynthesis was not affected by oxygen conditions. Dashed lines and solid lines indicate multiple steps and one metabolic step(s) needed to obtain the metabolite, respectively. Error bars, means Ϯ standard deviations (SD) (n ϭ 3 bacterial cultures per group); TCA, citric acid cycle; PPP, pentose phosphate pathway.
metabolites, 13 C incorporation rates were different under different oxygen conditions. For pyruvate, alanine, valine, and acetate, greater label ratios were observed under anaerobic conditions at 24 h. In contrast, lactate, glycine, serine, and isoleucine had greater label ratios under aerobic conditions at 24 h. The incorporation of [U-13 C 6 ]glucose into leucine biosynthesis was not impacted by oxygen conditions. Carbon fate in R. mucilaginosa diverged after 3-phosphoglycerate. The 13 C label was incorporated into serine and glycine, or into pyruvate, the precursor for lactate, acetate, and some amino acids.
Cross-feeding interactions between R. mucilaginosa and P. aeruginosa. In order to study the impact of R. mucilaginosa metabolites on P. aeruginosa, we cross-fed supernatant from an aerobic 48-h R. mucilaginosa culture to P. aeruginosa grown under low-nutrient conditions (M9 minimal medium) and nutrient-rich conditions (artificialsputum medium). P. aeruginosa was grown for 120 h before the cells were harvested in order to recapitulate the low growth rates of bacteria in CF patient sputa (10). The R. mucilaginosa supernatant included labeled lactate, pyruvate, and alanine ( Fig. 2A;  Fig. S1; Table S2). P. aeruginosa utilized R. mucilaginosa-derived metabolites to produce metabolites in M9 minimal medium and artificial-sputum medium. For example, although labeled lactate was found in the R. mucilaginosa supernatant, it was not detected in P. aeruginosa cultures, suggesting that P. aeruginosa consumed R. mucilaginosa-derived lactate ( Fig. 2A; Fig. S1; Table S2). P. aeruginosa utilization of lactate and other fermentation products has been observed in other studies (6,13). Since lactate levels have been reported as an indicator of CF patient response to antibiotic therapy, the finding that P. aeruginosa consumes lactate derived from another CF microbe may have clinical implications (14).
Labeled metabolites detected in P. aeruginosa cells grown in minimal medium included pyruvate, alanine, valine, serine, glycine, leucine, and isoleucine ( Fig. 2A;  Fig. S1; Table S2). In addition, isotope enrichment for serine, glycine, leucine, and isoleucine was greater in P. aeruginosa cells than in the supernatant of R. mucilaginosa, indicating that P. aeruginosa biosynthesized those metabolites. In contrast, when P. aeruginosa was grown in artificial-sputum medium, P. aeruginosa had higher levels of a single isotope-enriched amino acid (isoleucine) than occurred in the R. mucilaginosa supernatant ( Fig. 2A; Fig. S1; Table S2). Interestingly, although the R. mucilaginosa supernatant contained only unlabeled glutamate ( Fig. 2B and C; Fig. S2; Table S3), labeled glutamate was detected in both P. aeruginosa cultures (Fig. 2B, D, and E; Fig. S2; Table S3). This suggests that P. aeruginosa biosynthesized glutamate from 13 C sources in the R. mucilaginosa supernatant even in a nutrient-rich background with initially freely available glutamate (Text S1).
Glutamate provides a link between nitrogen and carbon metabolism by serving as a major amine group donor in transamination reactions for the synthesis of additional amino acids and nucleosides. In Escherichia coli, up to 88% of the total nitrogen that ends up in a biomass comes from glutamate, and the cellular glutamate pool needs to be kept high to drive the transamination reactions (15). In P. aeruginosa specifically, glutamate is a component of the cell wall and may play a role in P. aeruginosa virulence (16). Glutamate enhanced the yield of a virulence factor, exotoxin A (17), and induced swarming motility in P. aeruginosa on semisolid surfaces (18). More recently, glutamateinduced dispersion via c-di-GMP signaling pathways has been suggested (19). Glutamate might be derived from glutamine or alpha-ketoglutarate (20)(21)(22). However, the abundance of these two compounds was below the limit of quantification in this study. Future studies are needed to examine the biosynthesis pathways of glutamate and its role in the metabolism and physiology of P. aeruginosa. In summary, this study provides evidence that metabolite cross-feeding exists between R. mucilaginosa and P. aeruginosa, two common microorganisms found in polymicrobial communities in CF patient airways. The results from our study provide evidence that the physiology of CF pathogens can be influenced by the metabolic capabilities of other nearby microor-ganisms, even in a nutrient-rich environment, which can be tracked with stableisotope-labeled metabolomics.
Culture conditions and metabolomics. The bacterial strains chosen for this study were isolated from CF patients at the UCSD Adult CF Clinic: Pseudomonas aeruginosa PaFLR01 and Rothia mucilaginosa RmFLR01 (11,23). First, we took time points from R. mucilaginosa cultures to examine the kinetics of metabolites in glycolysis, the TCA cycle, amino acid biosynthesis, short-and long-chain fatty acid biosynthesis, and the pentose phosphate pathway in R. mucilaginosa, which was grown in triplicate in artificial-sputum medium (24) spiked with 100 mM [U-13 C 6 ]D-glucose (Sigma-Aldrich and Cambridge Isotope Laboratory) under anaerobic and aerobic oxygen conditions (5% CO 2 ) at 37°C. R. mucilaginosa cells were harvested at 4 h, 8 h, and 24 h. For the metabolite cross-feeding study, R. mucilaginosa was grown in the same medium aerobically for 48 h. The R. mucilaginosa supernatant was collected by filtering the culture, and the supernatant was diluted 10-fold in M9 minimal medium supplemented with succinate and in fresh artificial-sputum medium. P. aeruginosa was grown in triplicate aerobically, and the cells were harvested at 120 h. Metabolite extraction and data acquisition were carried out by following West Coast Metabolomics Center standard operating procedures (Text S1). Agilent MassHunter quantitative analysis software (v. B.07.00) was used for raw data processing. Natural abundance was corrected when isotope enrichment was calculated.

ACKNOWLEDGMENTS
This work was supported as a pilot project by the NIH (grant DK097154). K.L.W. is supported by a Gilead CF Research Scholars Award (app_00b072). T.G. is supported through the National Science Foundation's Integrative Graduate Education and Research Traineeship (IGERT) program (grant DGE-1144901).